Reliability of self-sampling for accurate assessment of respiratory virus viral and immunologic kinetics

Alpana Waghmare; Elizabeth M Krantz; Subhasish Baral; Emma Vasquez; Tillie Loeffelholz; E. Lisa Chung; Urvashi Pandey; Jane Kuypers; Elizabeth R Duke; Keith R Jerome; Alexander L Greninger; Daniel B Reeves; Florian Hladik; E Fabian Cardozo-Ojeda; Michael Boeckh; Joshua T Schiffer

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.03.20051706 doi: medRxiv preprint granzyme B, perforin, IL-6, IL-1α, MIP-1α and IFNγ all had high viral loads. These samples were 190 from two participants. All six participants had some samples in the least inflammatory class 191 (grey) and 5 participants had samples in the moderate inflammatory class (green). These data 192 indicate that the inflammatory immune milieu in the HRV-infected nasal passage is dynamic 193 over time, but tilts toward higher inflammation with higher viral loads. Very similar results were 194 observed when all samples were analyzed though only two classes of samples were 195 distinguished (Supp Fig 3b) . IFNγ. In the case of these viruses, the more inflammatory cytokine cluster clearly associated 201 with high viral loads for both RSV and MPV (Fig 3d) . 202 203 Mathematical modeling. We performed mathematical modeling separately on data from the 204 participant infected with RSV to examine whether complex immune and viral data from our 205 samples could be coupled mechanistically. We first developed the ordinary differential equation 206 model in equation (1) to link RSV viral load and early and late immune responses and evaluated 207 which cytokines may track those responses. For the early immune response, we found that only 208 the log 10 of the concentration change of IFN-γ and IP-10 was positively correlated to the viral 209 load during the first 5 days after enrollment ( Supp Fig 4a-b) , so we evaluated models for only 210 these two cytokines to track the RSV-early immune response. For the late immune responses, 211 we evaluated the model for fit to all observed cytokines (Supp Fig 4c) . An equivalent approach 212